Multiplicative vs additive seasonality

The first thing you need to decide is whether to apply transformations to the time series. In these last few exercises you will be working towards making a forecast of the CO\(_2\) time series. This and another time series for electricity production are plotted below.

Which of the above time series, if either, should be log transformed?

This exercise is part of the course

ARIMA Models in Python

Hands-on interactive exercise

Turn theory into action with one of our interactive exercises

This exercise is part of the course

ARIMA Models in Python

AdvancedSkill Level

4.8+

Start Course for Free

Dive straight in and learn about the most important properties of time series. You'll learn about stationarity and how this is important for ARMA models. You'll learn how to test for stationarity by eye and with a standard statistical test. Finally, you'll learn the basic structure of ARMA models and use this to generate some ARMA data and fit an ARMA model.

Exercise 1: Intro to time series and stationarity Exercise 2: Exploration Exercise 3: Train-test splits Exercise 4: Is it stationary Exercise 5: Making time series stationary Exercise 6: Augmented Dicky-Fuller Exercise 7: Taking the difference Exercise 8: Other tranforms Exercise 9: Intro to AR, MA and ARMA models Exercise 10: Model order Exercise 11: Generating ARMA data Exercise 12: Fitting Prelude

What lies ahead in this chapter is you predicting what lies ahead in your data. You'll learn how to use the elegant statsmodels package to fit ARMA, ARIMA and ARMAX models. Then you'll use your models to predict the uncertain future of stock prices!

Exercise 1: Fitting time series models Exercise 2: Fitting AR and MA models Exercise 3: Fitting an ARMA model Exercise 4: Fitting an ARMAX model Exercise 5: Forecasting Exercise 6: Generating one-step-ahead predictions Exercise 7: Plotting one-step-ahead predictions Exercise 8: Generating dynamic forecasts Exercise 9: Plotting dynamic forecasts Exercise 10: Intro to ARIMA models Exercise 11: Differencing and fitting ARMA Exercise 12: Unrolling ARMA forecast Exercise 13: Fitting an ARIMA model Exercise 14: Choosing ARIMA model

In this chapter, you will become a modeler of discerning taste. You'll learn how to identify promising model orders from the data itself, then, once the most promising models have been trained, you'll learn how to choose the best model from this fitted selection. You'll also learn a great framework for structuring your time series projects.

Exercise 1: Intro to ACF and PACF Exercise 2: AR or MA Exercise 3: Order of earthquakes Exercise 4: Intro to AIC and BIC Exercise 5: Searching over model order Exercise 6: Choosing order with AIC and BIC Exercise 7: AIC and BIC vs ACF and PACF Exercise 8: Model diagnostics Exercise 9: Mean absolute error Exercise 10: Diagnostic summary statistics Exercise 11: Plot diagnostics Exercise 12: Box-Jenkins method Exercise 13: Identification Exercise 14: Identification II Exercise 15: Estimation Exercise 16: Diagnostics

In this final chapter, you'll learn how to use seasonal ARIMA models to fit more complex data. You'll learn how to decompose this data into seasonal and non-seasonal parts and then you'll get the chance to utilize all your ARIMA tools on one last global forecast challenge.

Exercise 1: Seasonal time series Exercise 2: Seasonal decompose Exercise 3: Seasonal ACF and PACF Exercise 4: SARIMA models Exercise 5: Fitting SARIMA models Exercise 6: Choosing SARIMA order Exercise 7: SARIMA vs ARIMA forecasts Exercise 8: Automation and saving Exercise 9: Automated model selection Exercise 10: Saving and updating models Exercise 11: SARIMA and Box-Jenkins Exercise 12: Multiplicative vs additive seasonality

Current Exercise

Exercise 13: SARIMA model diagnostics Exercise 14: SARIMA forecast Exercise 15: Congratulations!